CanRouter AI
One endpoint · Every model · Transparent billing

Every model. One API.

CanRouter gives your application a single OpenAI-compatible endpoint to the best AI models — cheap, per-token pricing with one key, one credit balance, and full usage analytics built in.

No subscription. Pay only for what you call.

A model for every workload

Browse categories, not vendor silos. Pick the right model per task and switch freely without new integrations.

Programming

Code generation, review, refactoring, and agent workflows.

Roleplay

Creative writing, conversation, and character-driven chat.

Marketing

Copywriting, SEO content, and campaign strategy.

Science

Research, data analysis, and technical reasoning.

Translation

Multilingual translation and localization.

Image generation

Text-to-image across leading generative models.

Video generation

Text-to-video and video editing models.

Audio & speech

Speech-to-text, text-to-speech, and audio understanding.

Live in four steps

1

Create an account

Sign up in under a minute. No credit card required to explore.

2

Add credits

Top up once and use the balance across every model.

3

Generate an API key

One key, shown once, stored hashed. Rotate it any time.

4

Make requests

Point your existing OpenAI-compatible client at our endpoint.

Drop-in from day one

Keep your existing SDKs and tooling. Change the base URL, set one key, and every request flows through CanRouter with automatic routing, metering, and monitoring.

  • ✓ Streaming and non-streaming chat completions
  • ✓ Embeddings for retrieval workloads
  • ✓ Per-request usage and cost metadata
  • ✓ Request IDs you can trace end to end
curl https://api.canrouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $CANROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "canrouter-core",
    "messages": [
      { "role": "user", "content": "Explain rate limiting in one sentence." }
    ]
  }'

Transparent, per-token pricing

Every model lists its exact input and output price. Your dashboard shows the charge for every request — no surprises.

Frequently asked questions

Is the API OpenAI-compatible?

Yes. Our chat completions endpoint accepts the familiar request and response shapes, so SDKs and tools that speak the OpenAI protocol generally work with only a base-URL change.

How does billing work?

You keep a prepaid credit balance. Each request is priced per model at the moment it runs, and the exact charge is recorded on your ledger — visible in real time in the dashboard.

Do you store my prompts and responses?

No. We record usage metadata only — token counts, cost, latency, and status. Prompt and response bodies are never stored by default.

What happens when my balance runs out?

Requests are rejected with a clear billing error before reaching the provider. Nothing is retried, and you are never charged for a rejected request.

Can I switch models without changing code?

Yes — change the model field in your request. Provider routing happens on our side, so your integration stays the same.

Is CanRouter the cheapest way to access AI models?

Our catalog is repriced against major marketplaces twice a day, and a dedicated discounted tier runs below standard list prices — for many popular LLM, image, video, and audio models, CanRouter is the cheapest OpenAI-compatible access available.

Ready to route everything through one API?

Create an account, add credits, and ship in minutes.

Compare CanRouter's AI API pricing with OpenRouter, Haimaker, Together AI, Groq, Fireworks AI and other providers — one key, every model, one bill.